__tinygrad__ @__tinygrad__
Social handles Browse matching posts and comments already imported to Moot, including those shared before signup.
Activity Posts Comments Followers We're hiring a head of supply chain in Hong Kong. Must have knowledge of building computers, speak Mandarin, be good at operations, and be ready to work in person in Sheung Wan.
E-mail george@tinygrad.org to apply. Good application: come to the office for an interview next week.
#hiring #supply chain #tinygrad
your agent can't read 10 million lines and spend 30 minutes rebuilding large c projects, but it can read 26k and wait no time modding pure python. it will optimize everything from the model runner to the kernel to the driver, so of course it's the fastest. tinygrad will win.
E ethan thoma @EthanBThoma X Best part about frontier models is that i can use tinygrad for my all my experimental work and get codex/claude to make the kernels faster than pytorch/jax
9d ago View on X #tinygrad #machine learning #code optimization
Added some UIR examples to the course website. You are going to be spending a lot of time with this over the 14 weeks. By week 6 you'll have clean MNIST trainers and Qwen runners written in UIR.
#uir #qwen #mnist #ai course
tinybox red 4xMI300X. Batch 1 preorder, 25 units. $60k.
#localllama #mi300x #llm #benchmark
In a few more years it should be far ahead of everything. If Linux/LLVM is the stack for software 1.0, tinygrad is the stack for software 2.0.
B brrrock @bchap1n X @CerebralCoding_ what tinygrad is doing for amd software is so far ahead of anything on macOS
12d ago View on X #tinygrad #machine learning #deep learning #software 2.0 #ai frameworks #llvm
Course website is live.
#course announcement #university policy #in-person classes #online education #student accessibility #machine learning
I've been using Claude with pi, and they injected some trick to make the session keep retrying and failing to compact. Never hit this before. Full weekly usage used by one session in an hour. This is the real reason why the 5 hour limit was removed, user hostile beyond belief.
Claude @claudeai X We’re renaming our Max plans:
Max 5x → Max
Max 20x → Ultra
Max is still 100 / m o n t h , a n d U l t r a i s s t i l l 100/month, and Ultra is still 100/ m o n t h , an d U l t r ai ss t i l l 200. If you’re a subscriber, there’s nothing you need to do.
We’re also removing the 5-hour limit for verified Ultra accounts, starting today with most existing subscribers. Newer accounts keep a temporary limit until they’re verified, which helps us prevent fraud and abuse. You'll get an email when it's removed.
Weekly limits remain unchanged. View our plans here: http://claude.com/pricing
17d ago View on X #claude ai #usage limits #llm
If you are a CMU student signing up for 10-411, come join the tinygrad discord and say hi in the compilers-class channel.
#compilers #programming
can I interest you in an mnist traininer? #machine learning #mnist #deep learning #course project #neural networks
And as predicted, it happened! Price of R9700 has increased. 19 k y e s t e r d a y , 19k yesterday, 19 k y es t er d a y , 22k today. Will you buy before it's $25k?
#crypto #altcoin #price prediction
GLM-5.3-Flash now at 3200+ tok/s prefill, 155+ tok/s interactive decode on a $19k tinybox red flash. Get yours before prices go up!
#glm-5.3-flash #llm inference #ai hardware #benchmarking
Now benchmarked at 3200 tok/s prefill and 220 tok/s decode on a $19k base model tinybox red flash. Test API available to everyone with access to our network. More benchmarks incoming. This deal will not last!
the tiny corp @__tinygrad__ X GLM-5.3-Flash now at 3200+ tok/s prefill, 155+ tok/s interactive decode on a $19k tinybox red flash. Get yours before prices go up!
29d ago View on X #llm #inference #benchmark
It's better than an RTX 6000 for a third of the price. #graphics cards #rtx 6000 #gpu benchmarks #price-performance
We're testing a prototype #tinygrad #prototype #deep learning #programming